KTH, a Swedish based research institute, announced that it is working with Wikipedia to develop a crowd sourced speech synthesis platform which will be used on the Wikipedia website. While this text to speech engine will be developed to work optimally with Wikipedia, it will be open source, and expected to work with any website developed that uses the Wikimedia software. Open source means that anyone is able to freely develop plug-ins which work with other languages, or even screen readers.
3 languages will be supported from the start, English, Swedish, and basic Arabic. It is projected to be ready for launch in September of next year.Source: KCH
This made my day. Not only do we need a genuinely good open source speech engine other than Espeak, but there are precious few good Swedish TTS on the market either. Acapela's the best one I've found but it, obviously, isn't free let alone open source. Here's hoping this means we'll see a new, better default voice distributed with NVDA sometime in the near future. It's always the number one thing people talk about when I show them NVDA. They hate the voice and I don't blame them as it gives me headaches if I listen to Espeak too long. Here's hoping this project will bare a ripe fruit.
You must be logged in to post comments.